Bias range and selenium rectifier

[mrshow4u]

I saw a lot of stuff on Dyna Stereo-70, but didn't find anything specifically regarding this. I'm working on a Stereo 70 for a freind. The EL-34's are tired. ...and it won't bias properly.

So I've got a couple of questions:
1. Does anybody know the Vf of the original selenium rectifier?
2. Has the bias requirement for EL-34's varied widely over the years?
3. Do EL-34's still operate at idle normally, even though their output power capability has diminished?

This amp has the Selenium rectifier replaced single junction silicon.
Both channel put out about 10 Watts.
The plates are up at 425 VDC
Both channels won't bias, both under biased (30mA per tube rather than 50mA per tube)

The bias controls are shoved into a divider with a 3mA current. The bias range is -30V to -45V. I'd like to increase the current of the divider to get the the max bias setting to around -26V and retain the min setting of -45V. ....all the while keeping the stock bias controls. It looks like I can bump the divider current to 3.8mA, and get my range and be safe dissapation-wize.

Does anybody have any knowledge of what kind of variabilty exist in tubes these days and what sort of bias voltage range I should design for (with 425 plate voltage)?

Thanks, Jon

I searched the archives and found a lot of stuff on Stereo 70's, EL-34's, but didn't find anything specific for me. Apologies if the topic has been well covered.

[DigitalJunkie]

Replace that selenium!!

WHEN it fails (not IF) it will stink realllly bad,and cook the output tubes from lack of bias.

Replace the selenium rectifier with a 1N4007 or similar.You can tweak the overall voltage range by adjusting the values of the 10K resistors.
(you might need to,due to the lower forward drop of silicon vs. selenim)

Triode Electronics used to have a kit that did exactly that..
That was the first mod to my ST70.. No cooked tubes for me!

[mrshow4u]

Hi DigitalJunkie, thanks for the reply.

The Selenium has already been replaced. Not by me, but I would have done it. I don't even have the old Selenium rectifier (soon to morph into a capacitor) around. I kind of wanted to know what the Vf was on the Selenium, so I could make the bias circuit "thought" it was being fed by the Selenium. I just wanted to have the same bias range.

...at least to start with. Then I thought I might open up the bias range if it was felt that the variance in tubes these days demanded a greater bias range. I seem to remember certain batches of EL-34's couldn't be properly biased, even with a normal grid voltage.

[gingertube]

I can't recall exactly what the typical voltage drop across a Selenium rectifier was - it depended upon how many "stacks" were used which in turn depended upon the withstand voltage required. That is, various selenium rectifiers had various voltage drops.

Suffice to say that it was a lot more than a silicon diode and therein lies your problem. The rectifier diode / capacitor feeds the top off the bias pot. The bottom of the bias pot goes to 0v via a 10K?? resistor.

Because you now have more volts at the top of the pot you can't turn it down enough to get the tube idle currents high enough.

From your description the bias supply is now at -45V. The range you can set is -45V to -30V and you have 3mA down through the pot (so you have a 5K pot and a 10K resistor from the bottom of the pot to 0V - right?)

If you want to be able to turn down to -25V say (without changing that 10K) then you need to reduce the overall current to 2.5mA which is what it probably was when the selenium rectifier was fitted.

45V / 2.5mA = 18K total resistance so add a 3K3 resistor between the bias supply and the top off the pot.

Actually a bit of back calculation tells us that there was probably about 7 to 8V drop in that selenium rectifier and that the original bias supply voltage was about -37 volts.

There some basic Ohms law.
Hope this helps.
Cheers,
Ian

[pmillett]

Have a look at this:

http://www.pmillett.com/file_downloads/selenium.pdf

I didn't look carefully but the data on the last page implies a forward drop around ~30V for a bridge rated for 48VAC input voltage.

Pete

[mrshow4u]

Hi gingertuge, Thanks for the reply.
Yes you are correct regarding the 10k resistor and all in the circuit. I realize there will be shift in the range when moving to the Silicon rectifier. Yes, it is an option to model the Selenium rectifier through a resistor change. What I'm not sure about is, if I get back to the original bias range, will that be suitable for today's tubes, or has tube manufacturing standards created a wider variance of bias requirements?? I'm all over ohms law and getting the voltages I need. I don't want my freind to come back to me and say "Hey I just got these new tubes, but I can't bias them". I remember running into that with a McIntosh and Sovtek 6L6's. I think I'd like to put a greater bias range in using the same 5k bias pots. This will mean a little higher current and a little more power dissipated by the pots and the divider, but it looks like the margin is good.

Hi pmillett, Thank you for your reply.
Thanks for the rectifier link. I googled like mad, but found very little regarding Vf per plate and condstruction details. That is very helpful. Thank you.

[kevinkr]

I seem to recall that adding something like a 330 ohm resistor in series with the diode works well, might need to adjust up or down slightly.
With modern El34 you probably want an adjustment range of about -28 to -42V.

Incidentally the 425V you report for B+ is way too low, it should be somewhere between 475V - 500V under typical line voltage conditions. I think you probably need a new quad cap.

The low B+ will also play a role in your inability to bias the amplifier correctly. EL34 grid bias for 50mA in a properly working ST-70 will typically range from about -32V to -36V depending on brand. YMMV

[mrshow4u]

Hi Kevinkr, I'll check the filter sections. Thanks for the info on typical EL34 bias ranges too. I'll set the circuit up apprpriately.

Thanks again,

Jon

[anatech]

Hi Jon,
Just to throw another viewpoint. Selenium rectifiers have high bulk resistance, so the voltage drop varies a lot with current draw. I have a box of new green ones and have replaced plenty over the years. Normally, by the time you see a selenium rectifier these days, it is already supplying a lower voltage than it did new. I always end up using a silicon diode with a series resistor and messing around until the voltages agree with the schematic.

Biasing specs for tubes have not changed over the years, expect some brand to brand differences. With your old ones, I am not surprised they aren't biasing. Once you fix the bias network, make sure those are in at the bottom of the range. Check with some new ones. High bias is far worse than low bias for tube life, and yours are admittedly old.

-Chris

[mrshow4u]

Thanks Chris, so Seleniums are "squishy"eh? Understood about under vs over biasing. Yeah when the outputs are orange and it ain't from the heater.... It's okay it creates an OPPORTUNITY (marketing speak) to try new tubes. I've got current monitoring on my VariAc. I'm careful with the magic smoke. I've still got to de-bug the chassis, then debug the circuit. I'm starting to think that that 425V plate is telling me that one of the filter sections is opened. I hate that because it's probably 75% okay and it's hard to find 500WVDC caps. I'll tack the replacemants underneath if a sections open. (ugh) Let the Kluging begin.

Thanks